Dwarf Fortress Bug Tracker - Dwarf Fortress
View Issue Details
IDProjectCategoryView StatusDate SubmittedLast Update
0002463Dwarf FortressWorld Generation -- Generalpublic2010-06-24 16:552012-02-20 14:43
Reporterepsilon 
Assigned ToFootkerchief 
PrioritynormalSeverityminorReproducibilityalways
StatusresolvedResolutionduplicate 
PlatformPCOSWindows XPOS VersionProfessional
Product Version0.31.08 
Target VersionFixed in Version 
Summary0002463: Reproducible crash on worldgen
DescriptionAfter generating a world, browsing around it, then aborting and going to create a new world, DF crashes. Happens on medium (129x129) regions at the -very- least. The crash seems to happen between cave civilizations and regular civilizations. Also of note is that it didn't happen with a pocket region.
Steps To Reproduce1. Create a new "MEDIUM REGION"
2. Start a new game in Dwarf Mode
3. Browse across the entire map - pattern doesn't appear to matter, but getting all around the map does seem to matter.
4. Escape out, abort game
5. Create a new "MEDIUM REGION"
6. Crash before civ placement
Additional InformationGoing all around the map -is- important to reproducing the bug. The pattern doesn't seem to matter, but I have a habit of looking all over the map for that "perfect embark spot" - and I quit out when I can't find it, formerly by closing the process. Since it happens with a medium region and not with a pocket region, I'd -guess- that it's a memory problem of some sort. :p
Tags0.31.08, civilization, crash, worldgen
Relationships
duplicate of 0005077resolved Toady One Browsing world gen map causes corrupted feature files 
related to 0000046resolved Footkerchief After "Create New World Now!", there was a memory leak and process left running after quitting. 
Attached Files
Issue History
Date ModifiedUsernameFieldChange
2010-06-24 16:55epsilonNew Issue
2010-06-24 16:59epsilonNote Added: 0009064
2010-06-24 17:04epsilonNote Edited: 0009064bug_revision_view_page.php?bugnote_id=0009064#r3469
2010-06-24 17:05epsilonTag Attached: 0.31.08
2010-06-24 17:05epsilonTag Attached: civilization
2010-06-24 17:05epsilonTag Attached: crash
2010-06-24 17:05epsilonTag Attached: worldgen
2010-06-24 17:33snapNote Added: 0009065
2010-06-24 19:16Logical2uNote Added: 0009070
2010-06-24 19:16Logical2uTag Attached: AWAITING UPDATE
2010-06-24 20:13epsilonNote Added: 0009071
2010-06-24 20:14epsilonNote Edited: 0009071bug_revision_view_page.php?bugnote_id=0009071#r3476
2010-06-24 20:35epsilonNote Added: 0009073
2010-06-24 20:36epsilonNote Edited: 0009073bug_revision_view_page.php?bugnote_id=0009073#r3478
2010-06-24 20:37epsilonNote Edited: 0009064bug_revision_view_page.php?bugnote_id=0009064#r3479
2010-06-26 19:09epsilonNote Added: 0009154
2010-06-29 20:52FootkerchiefNote Edited: 0009065bug_revision_view_page.php?bugnote_id=0009065#r3585
2010-07-07 07:32FootkerchiefRelationship addedchild of 0002432
2010-07-07 07:36FootkerchiefRelationship addedrelated to 0000046
2010-07-07 07:36FootkerchiefRelationship deletedchild of 0002432
2010-07-13 21:55FootkerchiefRelationship addedrelated to 0002432
2010-07-13 21:55FootkerchiefTag Detached: AWAITING UPDATE
2010-07-13 21:57FootkerchiefNote Added: 0010200
2010-07-14 01:45epsilonNote Added: 0010225
2010-07-14 02:12epsilonNote Edited: 0010225bug_revision_view_page.php?bugnote_id=0010225#r3968
2010-07-14 02:18epsilonNote Edited: 0010225bug_revision_view_page.php?bugnote_id=0010225#r3969
2010-07-14 02:31epsilonNote Edited: 0010225bug_revision_view_page.php?bugnote_id=0010225#r3970
2010-07-25 13:48smjjamesNote Added: 0011005
2010-07-25 14:06FootkerchiefSummaryReproducable crash on worldgen => Reproducible crash on worldgen
2012-02-20 14:43FootkerchiefRelationship addedduplicate of 0005077
2012-02-20 14:43FootkerchiefStatusnew => resolved
2012-02-20 14:43FootkerchiefResolutionopen => duplicate
2012-02-20 14:43FootkerchiefAssigned To => Footkerchief
2012-02-20 14:44FootkerchiefRelationship deletedrelated to 0002432

Notes
(0009064)
epsilon   
2010-06-24 16:59   
(edited on: 2010-06-24 20:37)
Also of note is that it seems to work with an old region after a fresh start. You should be able to reproduce this yourself with any old medium region you've got hanging around.

Edit: Related to 0002432 - I've made that issue reproducible though. :D
(0009065)
snap   
2010-06-24 17:33   
(edited on: 2010-06-29 20:52)
Happens to me as well. Just made a large world, messed up a parameter and aborted to fix it. Crashed when making the second world got down to placing civilizations at about 14 (started at 50) and crashed out. Had it happen quite a few times already. Never had it crash on the first world generation after starting DF. Only if a previous world had been generated without closing/re-opening DF.
(0009070)
Logical2u   
2010-06-24 19:16   
What is your memory usage just prior to the crashes? And are you running a 32bit/x86 OS, and with how much memory?

I'm imagining this is just the "memory leak after world gen" bug that was previously reported.
(0009071)
epsilon   
2010-06-24 20:13   
(edited on: 2010-06-24 20:14)
32 bit OS, 3GB ram useable - but the memory usage didn't get over about 122MB on the task manager.
(0009073)
epsilon   
2010-06-24 20:35   
(edited on: 2010-06-24 20:36)
After some poking around, I'm assuming you're talking about issues 0000217 and 0000046. I'd believe it's the same issue, except that I generated a new world, searched around, and finally embarked on a 16x16 area with no issues (other than the thing being bloody slow) and no crashes.

If it were a memory leak, I'd think that it would be a bit more random than "reliably crashes after running the steps, and in the same spot each time" :D
(0009154)
epsilon   
2010-06-26 19:09   
Updating my info on this. I've been working with a small (62x62) region, and it will cause a crash with the steps if I browse the map from one end to the other - not even going up and down the map, just across it once.
(0010200)
Footkerchief   
2010-07-13 21:57   
Reminder sent to: epsilon

Did these crashes leave anything in your errorlog, e.g. "Impoverished Word Selector"?
(0010225)
epsilon   
2010-07-14 01:45   
(edited on: 2010-07-14 02:31)
Good question. I'll actually download .10 and recheck the whole mess. It was easy and reliable enough to reproduce the first time. I'll edit with an update within the hour.

Edit: Your update, sir.

Since I'm not entirely ready to play .10 yet, I ran the game straight from the "small" download - no music or sounds or whatever are included, but I didn't touch any of the init files either.

The result? Not a peep out of the error log. That is to say, the file wasn't generated at all. So no, no messages out of errorlog.txt

Edit of the edit:
http://www.bay12forums.com/smf/index.php?topic=33800.0 [^]

Looks like this has popped up before, minimal though the information may be.

Third edit:
I'm assuming that y'all, like most sane people, are asleep anyway and won't notice three edits. :P

A little quote from the DF2010 string dump at http://df.magmawiki.com/index.php/String_dump#Section_20 [^] :
"NAME_BUILDING_TEMPLE
NAME_ROAD
NAME_BRIDGE
NAME_TUNNEL
NAME_WALL
NAME_BUILDING_KEEP
*** Error(s) finalizing the symbol
Unrecognized word token:
*** Error(s) finalizing the translation
Impoverished Word Selector
Machine node out of bounds on split
Failed to find machine node frontier on split
Unreachable machine node on split"

If I had to guess, (and mind you, this is a VERY wild guess,) the "Impoverished Word Selector" error would seem to be linked to the choosing of names - for civilizations, dwarves, structures, etc. Assuming the info in the string dump hasn't been sorted in any way, it stands to reason that the last bits of the section with the error in it would be a part of the error check/sanity check routines in worldgen. Perhaps a better question would be, has it come up in places other than crashy worldgen, as implied by the post above?
(0011005)
smjjames   
2010-07-25 13:48   
Still happens in .12.

Also, I'm not getting any error about impovershed word selector with the crash. I am using the language files that come with Maydays Graphic pack, so that mught have something to do with the error not appearing.